Keywords

1 Introduction

Sheet metal molding is a technique whereby a human or machine introduces a metal mould into a thin metal sheet in order to transform the metal into the desired shape. In the context of automobile repair, this technique is utilized to return the automobile body back to its original shape. Each repair always bears some differences and may not be optimal. Therefore, a flexible approach should be taken to optimize the procedure under various circumstances. Furthermore, a significant amount of time is required to acquire an expert level of skill. Sheet metal molding for automotive repair is not sufficiently covered in vocational colleges and similar institutions. Therefore, non-experts must acquire skills through companies or union training, and textbooks outlining the process. As a result, each trainee develops their own and unique understanding of the process. In the context of declining birth rate, aging population, diminishing number of engineers, and retirement of many highly skilled individuals, it is feared that the current method of skill acquisition will be threatened by the shortage of experts in the near future. One way to address this issue is to utilize a motion analysis technique using a three-dimensional motion capture system. Motion analysis is already utilized in many forms of skill acquisition [1–8] and is regarded as an effective way of acquiring skills. Suzuki and Furuta clarified moving forms in hand sewing known as Unshin, a Japanese cultural skill, using a motion analysis computer system [1]. Murata and Iwase described the characteristics of baseball pitches from skilled pitchers by using a three-dimensional cinematographic analysis [2]. Hayashi and Yanagisawa showed skilled techniques for cutting cucumbers and carrots by using a video motion analysis system [3]. Digitization and characteristic evaluation of an expert’s motion were also performed using a three-dimensional motion measurement system for the following activities: a nail strike [4], burner’s work in glass processing [5], chest compression movement in lifesaving first aid [6], jump phase in a downhill race [7], and transfer care in elderly nursing [8] etc.

In this context, the objective of this study is to develop a sheet metalwork learning system for non-experts. In particular, the study focuses on clarifying the motion characteristics of experts when performing such a task. Experts and non-experts were filmed performing such a task and the differences in the process behavior and hammer striking position were examined.

2 Methodology

2.1 Outline of the Experiment

The fender of an automobile was repaired by workers by using a sheet metal working technique. The work was recorded using a video camera and a three dimensional motion measurement system. Then, the results were analyzed.

2.2 Experimental Facility

The experiment was carried out in March, 2015 at the Nara Auto Body Repair Association Training Facility. At the training facility, there was a booth with an automobile servicing lift. All measurements were carried out in this booth.

2.3 Participants

Participants were ten craftspeople involved in automobile repair in the Kinki Region. Those with ten or more years of experience were classified as experts and those with less than three years of experience were classified as non-experts. There were five expert participants and five non-expert participants. Table 1 presents the relevant details of the participants. Participants were informed that their work would be recorded and that the results of the analysis would immediately be made available to the public. Non-expert 5 had no work experience but did possess relevant background knowledge, working on a daily basis in the field of automotive paint repair.

Table 1. Participant information.
Full size table

2.4 Materials and Tools

Tools used in the experiment included a wooden hammer (hammer A, 213.3 g), three metal hammers (hammer B, C, D, weighing 433.4 g, 296.5 g, and 240.7 g, respectively), and three forming dollies (dolly A, B, C). Figure 1 displays these tools. The object of metalwork was the right front fender of a Corolla Fielder (Toyota Motor Corporation), shown in Fig. 2. The fender press line was dented with a heavy object (indicated by the red circle in Fig. 2), and the dent was defined as the object of repair.

Fig. 1.
figure 1

Hammers and dollies

Full size image
Fig. 2.
figure 2

Fender panel (Color figure online)

Full size image

2.5 Experimental Process

The participants were directed to repair the fender, which had been fixed to a real car, using the tools that we had prepared for them in advance. Participants were asked to repair the fender to a state that was ready for the next processing (puttying), and the decision on when this level had been reached were left up to them. However, it should be noted that a time limitation was imposed to finish the task such that the participants had to abandon the process if the task exceeded 30 min.

2.6 Recording Procedure

Infrared responsive markers were placed on 20 locations on each participant’s body, five of them on each hammer, and the other five on the fender, and then set as targets. A coordinate collection was performed with a MAC 3D System (Motion Analysis Corporation) equipped with an optical, three-dimension, and automated analysis device. The sampling frequency was set at 120 Hz. Figure 3 illustrates the recording setup. The frame of reference was set to the left-right movement of the participant on the x-axis, forward-backward movement on the y-axis, and up-down movement on the z-axis. Three video cameras were used to record the movement of the participant, concurrently.

Fig. 3.
figure 3

Recording setup

Full size image

2.7 Analysis

An analysis of the task was performed using the recorded video. In addition, the use of the hammer was also investigated. The position of the blow to the fender when using the hammer was measured through the three-dimensional motion analysis. The test was performed to investigate the differences between experts and non-experts. The level of significance was set at 5 %.

3 Results

3.1 Work Time

The task was divided into four processes: (1) Hammering using tools, (2) judging by hand, (3) judging by eyes, and (4) Others. Table 2 presents the work times for each process. The experts spent a significantly less amount of time compared to non-experts on hammering, others, and total work time, non-experts 2 and 3 were not able to complete the task during the allotted 30-min period. In contrast, experts 4 and 5 were able to complete the task within 5 min. The ‘others’ aspect of non-expert work time was long, and an examination of the details reveals that non-experts spent a significant amount of time on placing the dolly before hammering.

Table 2. Work time results
Full size table

Figure 4 shows the distribution of work processes for hammering, judging by hand, judging by eye, and others, totaling one hundred percent. Experts, as compared to non-experts, spent a smaller proportion of their time hammering and a greater proportion of their time checking their work by hand.

Fig. 4.
figure 4

Work rate (Color figure online)

Full size image

3.2 Hammering Time

Figure 5 shows the hammering time for hammers A through D. All experts used just one hammer to carry out the task. On the contrary, non-experts used two or three types of hammers. No experts chose to use the wooden hammer A.

Fig. 5.
figure 5

Working time by hammer (Color figure online)

Full size image

3.3 Strike Position

Figure 6 show the strike positions for expert 4 (the expert with the shortest hammering time), expert 1 (the expert with the longest hammering time), non-expert 1 (the non-expert with the shortest hammering time), and non-expert 2 (the non-expert with the longest hammering time), respectively. The black dots represent where the fender was struck, and the grey dots show the placement of the infrared markers. Compared to the strike positions of experts 1 and 4, the strike positions of non-experts 1 and 2 are broader in range. The same trend was observed for the other participants. The majority of strikes made by expert 4 were not on the press line where the dent was placed, showing that the repair work was performed by striking around the dented area.

Fig. 6.
figure 6

Strike position of (a) expert 4 (the expert with the shortest hammering time), (b) expert 1 (the expert with the longest hammering time), (c) non-expert 1 (the non-expert with the shortest hammering time), and (d) non-expert 2 (the non-expert with the longest hammering time)

Full size image

The hammer strike count is presented in Table 3. Expert 4, who had the shortest hammering time, also recorded the least number of strikes. Experts recorded a significantly lesser number of strikes compared to non-experts.

Table 3. Number of hammering strikes
Full size table

Figure 7 shows the relationship between hammering time and the number of strikes. There was no significant correlation between the expert and non-expert groups. However, there was a strong positive correlation for the participants as a whole group (r = 833, p = .00278) showing a clear trend where the number of strikes increased as the hammering time increased.

Fig. 7.
figure 7

Relationship between hammering time and strike count (Color figure online)

Full size image

4 Discussion

Excessive hammering stretches the metal, causing it to lose its characteristics. Non-experts who were not able to complete the task within 30 min did not suffer from time constraint. Instead, the dents that they created while attempting to repair the target dent also needed to be repaired. As a result, the number of areas to be repaired became too high and recovery was impossible. Although non-expert 1, who had the shortest hammering time among non-experts, had a hammering time of approximately 60 s, which is less than the time spent by expert 1 (the expert with the longest hammering time) the strike count of non-expert 1 was almost twice that of expert 1. Further, as shown in Fig. 7, there was a trend toward an increase in the number of strikes as hammering time increased. This observation indicates that spending more time on the task does not result in completion of repairing. The short repairing time taken by experts is because of their ability to concentrate one every blow of the hammer.

Figure 4 also shows that experts spent the majority of their time on validating their work, implying that they place great importance on this part. Differences in movement during the validation phase were apparent between experts and non-experts. Experts attempted to check the change of the fender in its entirety by standing up and changing their positioning. In contrast, non-experts focused only on the changes in front of them, the use of their hands was limited, and their field of vision did not extend to the whole fender.

5 Conclusions

With the objective of investigating the characteristics of how experts perform metalwork during an automotive repairing process, the work of experts and non-experts was analyzed. The results of the experiment indicate the following points:

  1. 1.

    The hammering time, others, and total work time of experts was significantly shorter than that of non-experts.

  2. 2.

    Experts completed the task by using only one hammer.

  3. 3.

    Experts used a significantly less number of strikes than non-experts, with the number of strikes increasing as the hammering time increased.